Getting the value: action steps

[MelSumner]

Breaking Issue #184 into action steps. Each of these items could be completed in distinct PRs, or a single PR.

• [ ] remove bullet points from step 2C
• [ ] re-word step 2C for improved clarity
• [ ] add example for typical use (we have one for complex use)
• [ ] propose a value computation algorithm (ensure to document exceptions/edge cases like contenteditable and lists)

[MelSumner]

Looking at my notes from different ACCVALUE discussions, after I condense them all, it seems like the biggest question is this:

Is there some implementor issue that means we need an accessible value computation algorithm?

---

Summary

There's the idea that obtaining an accessible value should be in the ACCNAME spec (Accessible Name, Description and Value). But what would an ACCVALUE calculation look like?

Considerations

Let's look at what already exists.

From the combobox documenation in ARIA spec

The value of a combobox is represented by one of the following:

If the combobox element is a host language element that provides a value, such as an HTML input element, 
the value of the combobox is the value of that element.

Otherwise, the value of the combobox is represented by its descendant elements and can be determined 
using the same method used to compute the name of a button from its descendant content.

It appears as though we already have an acceptable way to do this.

Using the same method to compute the name of a button from its descendant content

Questions

1. Is this question for more complex things like RTF? How should content in RTF be evaluated?
2. Do we need to come up with something different, or can we use what already exists in other specifications to get the value of a given input?
3. Would we expect that this would influence or change aria-valuenow?
4. Are there implementor issues with the method used to compute the name of a button from its descendant content?

[accdc]

We'll probably have to go over it again during one of the ARIA meetings, but from what I can recall when we spoke of this late last year in one of the other meetings, we determined that it wasn't necessary to create a whole new algorithm, but instead do the following.

1. Break out the value section within AccName where value is currently documented and move all of that into a new section for computing value within the spec.
2. Reference the new Value section within the AccName algorithm where applicable.
3. Update the different widget types within the Value section to better compute the value where needed on a case by case basis.

[spectranaut]

We talked about this briefly in the meeting last week, and Sarah said she remembered this issue and wanted to take a stab at it: https://www.w3.org/2024/02/22-aria-minutes#t04